Integrating Genetic Algorithm and Neural Network for the Reasons of Revisiting Frequency of Emergent Patients
黃堅展、陳郁文
E-mail: [email protected]
ABSTRACT
Data mining can find out hiden values of practical data for decision makers. Since the traffic condition is tense, and the patients of chronic and psychic disease become more and more, the emergency room（ER） is the busiest department in hospital. Improving the quality and capability of emergency room, in order to reduce the revisiting frequency and number of patients, is becoming an important issue of hospital. The data mining approach in this study is based on Back Propagation Neural Network, which integrating Genetic Algorithm（GA）to find the reasons of the revisiting frequency of patients in ER. We use the weights（as chromosome
）trained by neural network to form the initial population. After that, we calculate the fitness of each chromosome, the hit rate of each chromosome is defined as its fitness. After the evolution in GA, find the chromosome with the highest fitness, then find the reasons of revisiting frequency. This GA design can avoid local optimum in resolution and enhance the explanatory power of genetic algorithm / neural network for the revisiting frequency of emergent patients. Our studying results show that, the predicting accuracy of Genetic Algorithm Neural Network（GANN）, is significantly superior to only Back Propagation Neural Network（BPNN）.In addition, we use the relation matrix to find out the main reasons of influencing one-time revisiting. We found disposition, related clinic appointment, medical resource, temperature, pulse, breath are key factors, if the number of patients in emergency internal medicine ward exceeds 25 persons, the number of patients in surgery and trauma ward exceeds 20 persons and the number of patients in pediatrics ward exceeds 5 persons, then the medical quality will be worse. The aforementioned observations are valuable in practical ER services.
Keywords : Neural Network ; Genetic Algorithm ; Data Mining ; Emergent Patients Table of Contents
目 錄 封面內頁 簽名頁 博碩士論文暨電子檔案上網授權書... iii 中文摘要... iv ABSTRACT... v 誌謝... vi 目 錄... vii 圖目 錄... x 表目錄... xii 第一章 緒論... 1 1.1 研究背 景... 1 1.2 研究動機... 2 1.3 研究目的... 3 1.4 研究假設與限 制... 4 1.5 研究內容及流程... 5 第二章 文獻探討... 8 2.1 醫療品質指 標... 8 2.1.1 醫療品質指標定義... 8 2.1.2 台灣醫療品質指標計畫與台灣醫療照護品質指 標系 列... 9 2.2 資料探勘... 12 2.2.1 資料探勘的步驟... 13 2.2.2 資料探勘的技 術... 16 2.3 神經網路... 17 2.3.1 神經網路沿革與發展... 17 2.3.2 神經網路架 構... 19 2.3.3 倒傳遞神經網路... 26 2.4 基因演算法... 29 2.5 結合基因演算法與神 經網路的相關文獻... 35 2.6 關係矩陣... 38 2.7 小結... 39 第三章 研究方
法... 41 3.1 神經網路學習階段... 43 3.2 基因演算法調整階段... 51 3.3 神經網路回想 測試階段... 58 3.4 小結... 58 第四章 實例驗證與結果分析... 60 4.1 問題描
述... 60 4.2 研究變數屬性說明與編碼... 62 4.3 執行程式相關設定... 66 4.4 程式結果 比較與檢定... 71 4.5 敏感度分析... 75 4.6 關係矩陣運算... 78 4.7 回診原因討 論... 81 第五章 結論與建議... 102 5.1 結論... 102 5.2 建
議... 106 參考文獻... 109 附錄一 倒傳遞神經網路公式推導... 114 附錄二 程 式執行結果及說明... 120
REFERENCES
中文部分 [1] 王文俊，認識Fuzzy，全華科技圖書股份有限公司，2005。
[2] 王進德、蕭大全，類神經網路與模糊控制理論入門，全華科技圖書股份有限公司，2003。
[3] 台灣醫務管理學會，THIS交流園區，THIS指標明細。 http://www.tche.org.tw/this_data/this-index.htm [4] 吳肖琪，評估醫院醫療品 質指標，行政院研討會，1999。
[5] 李沃牆、林維垣，基因演化類神經網路模型於台股上現行權證的評價，東吳經濟學學術研討會論文集，2000。
[6] 沈希哲，由台灣醫療品質指標計畫急診指標探討急診醫療品質，TQIP研究計畫。 http://www.tjcha.org.tw/s_plans.asp?catid=48 [7] 周 鵬程，遺傳演算法原理與應用─活用Matlab，全華科技圖書 股份有限公司，2005。
[8] 周鵬程，類神經網路入門─活用Matlab，全華科技圖書股份有限公司，2002。
[9] 林建成，遺傳演化類神經網路於台灣股市預測與交易策略之研究，東吳大學經濟學研究所碩士論文，2002。
[10] 林靜怡，以系統動力學探討醫療品質指標系統之適切性─以急診品質指標為例，大葉大學事業經營研究所碩士論文， 2005。
[11] 林豐澤，演化式計算下篇:基因演算法以及三種應用實例，智慧科技與應用統計學報，第17卷第5期，第29-56頁，2003。
[12] 財團法人醫院評鑑暨醫療品質策進會，台灣醫療品質指標計畫，TQIP簡介。 http://www.tjcha.org.tw/quality.asp?catid=32 [13] 高全 興，類神經網路於空氣品質短期預測之研究，國立雲林科技大學環境與安全工程研究所碩士論文，1997。
[14] 陳清暉，類神經網路結合遺傳演算法應用於工業區聯合廢水處理廠放流水品質預測之研究，國立雲林科技大學環境與安全工程研究 所碩士論文，1999。
[15] 勞寬，醫院品質報告卡可行性研究，台北醫學大學醫務管理 學研究所碩士論文，2002。
[16] 楊東翰，整合基因演算法及類神經網路於印刷電路板生產預測之研究，元智大學工業工程與管理研究所碩士論文，2003。
[17] 楊朝旭、傅鍾仁，我國醫院服務品質指標認知差異之實證研究，第二屆服務業管理研討會，1996。
[18] 葉怡成，類神經網路模式應用與實作，儒林圖書有限公司，1995。
[19] 簡家宏，應用基因類神經網路於空氣品質短期預測及監測資料異常值診斷之研究－以台中縣沙鹿測站為例，國立雲林科技大學環境 與安全衛生工程研究所碩士論文，2004。
[20] 羅華強，類神經網路─MATLAB的應用，高立圖書有限公司，2005。
[21] 蘇木春、張孝德，機器學習:類神經網路、模糊系統以及基因演算法則，全華科技圖書股份有限公司，2002。 英文部分 [22]
Adriaans, P. and Zantinge, D., Data Mining, Addison-Wesley Publication Company, 1996.
[23] Agrawal, R., Imielinski, T., and Swami, A., “Database Mining: A performance perspective on Learning and Discovery in Knowledge-Based Database,” IEEE Trans on Knowledge and Data Engineering, Volume 5, No 6, pp. 914-925, 1993.
[24] Berry, M. and Gordon, S. L., Mastering Data Mining:The Art and Science of Customer Relationship Management, New York: John Wiley, 2001.
[25] Curt, H., “The Devil’s in the Detail: Techniques, Tools,and Application for Database mining and Knowledge Discovery-Part I,”
Intelligent Software Strategies, Volume 6, No 9, pp. 1-15, 1995.
[26] Donabedian, A., The quality of care:how can it be assessed?, JAMA, pp. 1743-1748, 1988.
[27] Fayyad, U.M., “Data Mining and knowledge Discovery: Making Sense Out of Data,” IEEE Expert, Volume 11, Issue 5, pp. 20-25, 1996.
[28] Fu Y., “Data mining task, technique and applications,” IEEE Potentials, Volume 5, pp. 18-20, 1997.
[29] Grupe, F. H., and Owrang, M. M.,“Data Base Mining Discovering New Knowledge and Cooperative Advantage,” Information System Management, Volume 12, No 4, pp. 26-31, 1995.
[30] Han, J. and Kamber, M., Data Mining: Concepts and Techniques, John Wiley & Son, 2001.
[31] Hand, D. J., Gordon, B., Kelly, M.G. and Adams, N. M., “Data mining for fun and profit,” Statistical Science, Volume 15, No 2, pp.
111-131, 2000.
[32] IBM, IBM Intelligent Miner for Data, 1998.
[33] Kim, S. M., Kim, J. D., Hong, J. H., Nam, D. W., Lee, D. H. and Lee, J. Y., A System for Association Rule Finding from an Internet Portal Site, 2000.
[34] Kim, K. J. and Han, I., “Genetic algorithms approach to feature discretization in artificial neural networks for the prediction of stock price index,” Expert Systems with Application, Volume 19, pp. 125-132, 2000.
[35] Kleissner, C., “Data mining for the enterprise,” In Proceedings of the Thirty-First Hawaii International Conference on, Volume 7, pp.
295-304, 1998.
[36] Linder, J. C., “outcomes measurement in Hospital:Can the system change the organization”, Hospital & Health Services Administration, volume 37, lssue 2, pp.143-166, 1992.
[37] Michael, J. A. and Linoff, G., Data Mining Technique: for Marketing, Sales and Customer Support, Wiley Computer Publishing, New York, 1997.
[38] Montana, D. and Davis, L., “Training feed forward neural networks using genetic algorithms,” Proceedings of 11th International Joint Conference on Artificial Intelligence, San Mateo, CA, Morgan Kaufmanns, volume 45, lssue 3, pp.762-767, 1989.
[39] Piatesky-Shapiro, G., Fayyad, U. M., Smyth P. and Uthurusamy, R., Advances in Knowledge Discovery and Data Mining, AAAI/MIT Press, 1996.
[40] Pyle, D., Data Preparation for Data Mining, Morgan Kaufmann, USA, 1999.
[41] Sexton, R.S. and Gupta, J. N., “Comparative evaluation of genetic algorithm and backpropagation for training neural networks,”
Information Sciences, Volume 129, pp.45-59, 2000.
[42] Srinivasan, D., “Evolving artificial neural networks for short term load forecasting,” Neural computing, Volume 23, pp.265-276, 1998.
[43] Ting, C. K., Li, S. T. and Lee, C. N., “TGA: A new integrated approach to evolutionary Algorithms,” Congress on Evolutionary Computation, Volume 28, pp. 917-924, 2001.
